I'm integrating Ory Cloud Kratos with my service h...
# ory-network
j
I'm integrating Ory Cloud Kratos with my service hosted in AWS. • I'm passing the URL I found under
Ory Kratos
to the GO generated client. • I set the Authorization header using the Personal Access Token • I verify the connection by calling
client.MetadataApi.GetVersion(ctx).Execute()
Connection verification fails with 404. It works in local development. When I visit the url i see an error:
Copy code
{
  "error": {
    "code": 404,
    "status": "Not Found",
    "request": "d2997dc1-7f49-981d-8217-51bb2be936a9",
    "message": "Requested url does not match any rules"
  }
}
h
ayyy 🙂 great to see you around! It’s quite late here, will answer tomorrow 🙂
1
j
I was never gone 😄
h
I don’t think we support that endpoint in Ory Cloud - it would give you the hash only more or less. I think we rewrote it to another endpoint for some internal use - do you need the version ready?
j
I don't really need the version, just a simple API to check connectivity. Configuring networking in AWS isn't that easy so I just want to periodically make sure that it works. Is there any other API that could be used for this purpose?
h
j
This API isn't in the client, but I see
health/alive
there, but requests like this don't work (I'll change the token later):
Copy code
curl  -H "Authorization:ory_pat_Yxl4F1PxTVk0byBocYCH3DO4SnC0x4kf" <https://dazzling-shtern-uw97dfb4yg.projects.oryapis.com/health/alive>

{"error":{"code":404,"status":"Not Found","request":"e6015336-e513-939d-9057-999007661e4e","message":"Requested url does not match any rules"}}
h
Yeah that endpoint doesn’t work in cloud unfortunately. The other endpoint is currently not documented 😕 you would need to use a native fetch
1
the idea is that the thing is always alive 😄 but I understand the need of a health check… do you need to check for kratos specifically? or just ory cloud in general?
j
I use only kratos for now
I'm checking for network fault on my side. AWS has a lot of configuration that can easily go wrong, I'd rather catch an error right away from log events
h
ahh i see that makes sense